🧠 globe_ai #

globe_ai is a Dart-first package for interacting with large language models (LLMs) like OpenAI’s GPT series.

✨ Features #

• 📝 generateText — basic prompt or messages based text generation

• 📡 streamText — stream text responses as they’re generated

• 🧱 generateObject — validate structured JSON output against a schema

• 🌊 streamObject — stream and validate structured data

🚀 Installation #

Install Globe Runtime #

You can install globe runtime using either of these approaches.

• Using Globe CLI.

  If you have Globe CLI installed locally, you can run globe runtime install.

• Adding Manually to your Environment

  Download the libglobe_runtime dynamic library for your Platform from Github Release Pages and place it in ~/.globe/runtime/ directory.

Setup #

• Add dependency

dependencies:
  globe_ai: ^<latest-version>

• Configure your model provider (e.g. OpenAI):

final model = openai.chat('gpt-4o', user: 'Chima');

or

final model = openai('gpt-4o');

Usage #

🔹 Text Generation #

final result = await generateText(
  model: openai.chat('gpt-4o'),
  prompt: 'What is the capital of Ghana?',
);
print(result);

or

final textWithPdf = await generateText(
  model: openai.chat('gpt-4o', user: 'Chima'),
  messages: [
    OpenAIMessage(
      role: 'user',
      content: [
        OpenAIInput(text: 'What is the title of this book?'),
        OpenAIInput(
          file: FileInput(
            data: File('bin/test_doc.pdf').readAsBytesSync(),
            mimeType: 'application/pdf',
            name: 'ai.pdf',
          ),
        ),
      ],
    ),
  ],
);
print(textWithPdf);

🔹 Streaming Text #

final stream = streamText(
  model: openai('o4-mini'),
  prompt: 'Describe the Mission burrito vs SF burrito debate.',
);

await for (final chunk in stream) {
  stdout.write(chunk);
}

🔹 Structured Object Generation #

final schema = l.schema({
  'recipe': l.schema({
    'name': l.string(),
    'ingredients': l.list(validators: [
      l.schema({'name': l.string(), 'amount': l.string()}),
    ]),
    'steps': l.list(validators: [l.string()]),
  })
});

final result = await generateObject<Map<String, dynamic>>(
  model: openai('gpt-4.1'),
  prompt: 'Generate a lasagna recipe.',
  schema: schema,
);

print(result['recipe']['name']);

🔹 Streaming Structured Objects #

final resultStream = streamObject<Map<String, dynamic>>(
  model: openai('gpt-4.1'),
  prompt: 'Generate a lasagna recipe.',
  schema: schema,
);

await for (final chunk in resultStream) {
  print(chunk); // Validated partial output
}

📚 Roadmap #

• 🌐 Outside-Globe support — Coming soon

• 🤖 Additional model providers — In progress

• 🧪 Unit tests & CI examples

• 📖 Function-level API docs

🛠️ Development #

• Building the JS package

  dart pub run rps build
  

• Generate types from protos

  dart pub run rps gen_dart
  

  and

  dart pub run rps gen_typescript